implemented workflow of creation action set for scholexplorer
parent
825d9f0289
commit
16c91203bd
@ -0,0 +1,90 @@
|
||||
package eu.dnetlib.dhp.sx.provision
|
||||
|
||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser
|
||||
import eu.dnetlib.dhp.schema.oaf.{Oaf, Relation, Result}
|
||||
import org.apache.spark.{SparkConf, sql}
|
||||
import org.apache.spark.sql.{Dataset, Encoder, Encoders, SaveMode, SparkSession}
|
||||
import org.slf4j.{Logger, LoggerFactory}
|
||||
|
||||
import scala.io.Source
|
||||
|
||||
object SparkCreateActionset {
|
||||
|
||||
def main(args: Array[String]): Unit = {
|
||||
val log: Logger = LoggerFactory.getLogger(getClass)
|
||||
val conf: SparkConf = new SparkConf()
|
||||
val parser = new ArgumentApplicationParser(Source.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/dhp/sx/actionset/generate_actionset.json")).mkString)
|
||||
parser.parseArgument(args)
|
||||
|
||||
|
||||
val spark: SparkSession =
|
||||
SparkSession
|
||||
.builder()
|
||||
.config(conf)
|
||||
.appName(getClass.getSimpleName)
|
||||
.master(parser.get("master")).getOrCreate()
|
||||
|
||||
|
||||
val sourcePath = parser.get("sourcePath")
|
||||
log.info(s"sourcePath -> $sourcePath")
|
||||
|
||||
val targetPath = parser.get("targetPath")
|
||||
log.info(s"targetPath -> $targetPath")
|
||||
|
||||
val workingDirFolder = parser.get("workingDirFolder")
|
||||
log.info(s"workingDirFolder -> $workingDirFolder")
|
||||
|
||||
implicit val oafEncoders:Encoder[Oaf] = Encoders.kryo[Oaf]
|
||||
implicit val resultEncoders:Encoder[Result] = Encoders.kryo[Result]
|
||||
implicit val relationEncoders:Encoder[Relation] = Encoders.kryo[Relation]
|
||||
|
||||
import spark.implicits._
|
||||
|
||||
val relation = spark.read.load(s"$sourcePath/relation").as[Relation]
|
||||
|
||||
relation.filter(r => (r.getDataInfo== null || r.getDataInfo.getDeletedbyinference == false) && !r.getRelClass.toLowerCase.contains("merge"))
|
||||
.flatMap(r => List(r.getSource,r.getTarget)).distinct().write.save(s"$workingDirFolder/id_relation")
|
||||
|
||||
|
||||
val idRelation = spark.read.load(s"$workingDirFolder/id_relation").as[String]
|
||||
|
||||
log.info("extract source and target Identifier involved in relations")
|
||||
|
||||
|
||||
log.info("save relation filtered")
|
||||
|
||||
relation.filter(r => (r.getDataInfo== null || r.getDataInfo.getDeletedbyinference == false) && !r.getRelClass.toLowerCase.contains("merge"))
|
||||
.write.mode(SaveMode.Overwrite).save(s"$workingDirFolder/actionSetOaf")
|
||||
|
||||
log.info("saving publication")
|
||||
|
||||
val publication:Dataset[(String, Result)] = spark.read.load(s"$sourcePath/publication").as[Result].map(p => (p.getId, p))
|
||||
|
||||
publication
|
||||
.joinWith(idRelation, publication("_1").equalTo(idRelation("value")))
|
||||
.map(p => p._1._2)
|
||||
.write.mode(SaveMode.Append).save(s"$workingDirFolder/actionSetOaf")
|
||||
|
||||
log.info("saving dataset")
|
||||
val dataset:Dataset[(String, Result)] = spark.read.load(s"$sourcePath/dataset").as[Result].map(p => (p.getId, p))
|
||||
dataset
|
||||
.joinWith(idRelation, publication("_1").equalTo(idRelation("value")))
|
||||
.map(p => p._1._2)
|
||||
.write.mode(SaveMode.Append).save(s"$workingDirFolder/actionSetOaf")
|
||||
|
||||
log.info("saving software")
|
||||
val software:Dataset[(String, Result)] = spark.read.load(s"$sourcePath/software").as[Result].map(p => (p.getId, p))
|
||||
software
|
||||
.joinWith(idRelation, publication("_1").equalTo(idRelation("value")))
|
||||
.map(p => p._1._2)
|
||||
.write.mode(SaveMode.Append).save(s"$workingDirFolder/actionSetOaf")
|
||||
|
||||
log.info("saving Other Research product")
|
||||
val orp:Dataset[(String, Result)] = spark.read.load(s"$sourcePath/otherresearchproduct").as[Result].map(p => (p.getId, p))
|
||||
orp
|
||||
.joinWith(idRelation, publication("_1").equalTo(idRelation("value")))
|
||||
.map(p => p._1._2)
|
||||
.write.mode(SaveMode.Append).save(s"$workingDirFolder/actionSetOaf")
|
||||
}
|
||||
|
||||
}
|
||||
@ -0,0 +1,86 @@
|
||||
package eu.dnetlib.dhp.sx.provision
|
||||
|
||||
import com.fasterxml.jackson.databind.ObjectMapper
|
||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser
|
||||
import eu.dnetlib.dhp.schema.action.AtomicAction
|
||||
import eu.dnetlib.dhp.schema.oaf.{Oaf, OtherResearchProduct, Publication, Relation, Software, Dataset => OafDataset}
|
||||
import org.apache.hadoop.io.Text
|
||||
import org.apache.hadoop.io.compress.GzipCodec
|
||||
import org.apache.hadoop.mapred.SequenceFileOutputFormat
|
||||
import org.apache.spark.SparkConf
|
||||
import org.apache.spark.sql.{Encoder, Encoders, SparkSession}
|
||||
import org.slf4j.{Logger, LoggerFactory}
|
||||
|
||||
import scala.io.Source
|
||||
|
||||
object SparkSaveActionSet {
|
||||
|
||||
|
||||
def toActionSet(item: Oaf): (String, String) = {
|
||||
val mapper = new ObjectMapper()
|
||||
|
||||
item match {
|
||||
case dataset: OafDataset =>
|
||||
val a: AtomicAction[OafDataset] = new AtomicAction[OafDataset]
|
||||
a.setClazz(classOf[OafDataset])
|
||||
a.setPayload(dataset)
|
||||
(dataset.getClass.getCanonicalName, mapper.writeValueAsString(a))
|
||||
case publication: Publication =>
|
||||
val a: AtomicAction[Publication] = new AtomicAction[Publication]
|
||||
a.setClazz(classOf[Publication])
|
||||
a.setPayload(publication)
|
||||
(publication.getClass.getCanonicalName, mapper.writeValueAsString(a))
|
||||
case software: Software =>
|
||||
val a: AtomicAction[Software] = new AtomicAction[Software]
|
||||
a.setClazz(classOf[Software])
|
||||
a.setPayload(software)
|
||||
(software.getClass.getCanonicalName, mapper.writeValueAsString(a))
|
||||
case orp: OtherResearchProduct =>
|
||||
val a: AtomicAction[OtherResearchProduct] = new AtomicAction[OtherResearchProduct]
|
||||
a.setClazz(classOf[OtherResearchProduct])
|
||||
a.setPayload(orp)
|
||||
(orp.getClass.getCanonicalName, mapper.writeValueAsString(a))
|
||||
|
||||
case relation: Relation =>
|
||||
val a: AtomicAction[Relation] = new AtomicAction[Relation]
|
||||
a.setClazz(classOf[Relation])
|
||||
a.setPayload(relation)
|
||||
(relation.getClass.getCanonicalName, mapper.writeValueAsString(a))
|
||||
case _ =>
|
||||
null
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
def main(args: Array[String]): Unit = {
|
||||
val log: Logger = LoggerFactory.getLogger(getClass)
|
||||
val conf: SparkConf = new SparkConf()
|
||||
val parser = new ArgumentApplicationParser(Source.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/dhp/sx/actionset/save_actionset.json")).mkString)
|
||||
parser.parseArgument(args)
|
||||
|
||||
|
||||
val spark: SparkSession =
|
||||
SparkSession
|
||||
.builder()
|
||||
.config(conf)
|
||||
.appName(getClass.getSimpleName)
|
||||
.master(parser.get("master")).getOrCreate()
|
||||
|
||||
|
||||
val sourcePath = parser.get("sourcePath")
|
||||
log.info(s"sourcePath -> $sourcePath")
|
||||
|
||||
val targetPath = parser.get("targetPath")
|
||||
log.info(s"targetPath -> $targetPath")
|
||||
|
||||
implicit val oafEncoders:Encoder[Oaf] = Encoders.kryo[Oaf]
|
||||
implicit val tEncoder:Encoder[(String,String)] = Encoders.tuple(Encoders.STRING,Encoders.STRING)
|
||||
|
||||
spark.read.load(sourcePath).as[Oaf]
|
||||
.map(o =>toActionSet(o))
|
||||
.filter(o => o!= null)
|
||||
.rdd.map(s => (new Text(s._1), new Text(s._2))).saveAsHadoopFile(s"$targetPath", classOf[Text], classOf[Text], classOf[SequenceFileOutputFormat[Text,Text]], classOf[GzipCodec])
|
||||
|
||||
}
|
||||
|
||||
}
|
||||
@ -0,0 +1,6 @@
|
||||
[
|
||||
{"paramName":"mt", "paramLongName":"master", "paramDescription": "should be local or yarn", "paramRequired": true},
|
||||
{"paramName":"s", "paramLongName":"sourcePath","paramDescription": "source path", "paramRequired": true},
|
||||
{"paramName":"w", "paramLongName":"workingDirFolder","paramDescription": "the working Dir Folder", "paramRequired": true},
|
||||
{"paramName":"t", "paramLongName":"targetPath","paramDescription": "the target path ", "paramRequired": true}
|
||||
]
|
||||
@ -0,0 +1,23 @@
|
||||
<configuration>
|
||||
<property>
|
||||
<name>jobTracker</name>
|
||||
<value>yarnRM</value>
|
||||
</property>
|
||||
<property>
|
||||
<name>nameNode</name>
|
||||
<value>hdfs://nameservice1</value>
|
||||
</property>
|
||||
<property>
|
||||
<name>oozie.use.system.libpath</name>
|
||||
<value>true</value>
|
||||
</property>
|
||||
<property>
|
||||
<name>oozie.action.sharelib.for.spark</name>
|
||||
<value>spark2</value>
|
||||
</property>
|
||||
|
||||
<property>
|
||||
<name>oozie.launcher.mapreduce.user.classpath.first</name>
|
||||
<value>true</value>
|
||||
</property>
|
||||
</configuration>
|
||||
@ -0,0 +1,76 @@
|
||||
<workflow-app name="Scholexplorer_to_ActionSet_Workflow" xmlns="uri:oozie:workflow:0.5">
|
||||
<parameters>
|
||||
<property>
|
||||
<name>sourcePath</name>
|
||||
<description>the path of the consistent graph</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>workingDirFolder</name>
|
||||
<description>the path of working dir ActionSet</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>outputPath</name>
|
||||
<description>the path of Scholexplorer ActionSet</description>
|
||||
</property>
|
||||
</parameters>
|
||||
|
||||
<start to="ExportDataset"/>
|
||||
|
||||
<kill name="Kill">
|
||||
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
||||
</kill>
|
||||
|
||||
<action name="createActionSet">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn-cluster</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Create Action Set</name>
|
||||
<class>eu.dnetlib.dhp.sx.provision.SparkCreateActionset</class>
|
||||
<jar>dhp-aggregation-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--driver-memory=${sparkDriverMemory}
|
||||
--conf spark.sql.shuffle.partitions=3840
|
||||
--conf spark.extraListeners=${spark2ExtraListeners}
|
||||
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
||||
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
||||
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
||||
</spark-opts>
|
||||
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
|
||||
<arg>--targetPath</arg><arg>${outputPath}</arg>
|
||||
<arg>--workingDirFolder</arg><arg>${workingDirFolder}</arg>
|
||||
<arg>--master</arg><arg>yarn-cluster</arg>
|
||||
</spark>
|
||||
<ok to="End"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
|
||||
<action name="SaveActionSet">
|
||||
<spark xmlns="uri:oozie:spark-action:0.2">
|
||||
<master>yarn-cluster</master>
|
||||
<mode>cluster</mode>
|
||||
<name>Save Action Set</name>
|
||||
<class>eu.dnetlib.dhp.sx.provision.SparkSaveActionSet</class>
|
||||
<jar>dhp-aggregation-${projectVersion}.jar</jar>
|
||||
<spark-opts>
|
||||
--executor-memory=${sparkExecutorMemory}
|
||||
--executor-cores=${sparkExecutorCores}
|
||||
--driver-memory=${sparkDriverMemory}
|
||||
--conf spark.sql.shuffle.partitions=3840
|
||||
--conf spark.extraListeners=${spark2ExtraListeners}
|
||||
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
||||
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
||||
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
||||
</spark-opts>
|
||||
<arg>--sourcePath</arg><arg>${workingDirFolder}/actionSetOaf</arg>
|
||||
<arg>--targetPath</arg><arg>${outputPath}</arg>
|
||||
<arg>--master</arg><arg>yarn-cluster</arg>
|
||||
</spark>
|
||||
<ok to="End"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
<end name="End"/>
|
||||
</workflow-app>
|
||||
@ -0,0 +1,5 @@
|
||||
[
|
||||
{"paramName":"mt", "paramLongName":"master", "paramDescription": "should be local or yarn", "paramRequired": true},
|
||||
{"paramName":"s", "paramLongName":"sourcePath","paramDescription": "source path", "paramRequired": true},
|
||||
{"paramName":"t", "paramLongName":"targetPath","paramDescription": "the target path ", "paramRequired": true}
|
||||
]
|
||||
Loading…
Reference in New Issue